Development, validation and application of a UPLC-MS/MS method for simultaneous quantification of OPC-61815 and its metabolites tolvaptan, DM-4103 and DM-4107 in human plasma.

OPC-61815 is an intravenous formulation vasopressin antagonist designed to treat heart failure patients, especially who have difficulty in oral intake. Tolvaptan together with DM-4103 and DM-4107 are considered as the major metabolites of OPC-61815 biotransformed in the liver via cytochrome P450 (CYP) 3A. An efficient and robust ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) method for quantification of OPC-61815 and its three metabolites in human plasma was developed and fully validated. To our best knowledge, it was the first published method that simultaneously quantified all of these four analytes in only one run. Simple and rapid sample preparation procedure and very short UPLC-MS/MS run time (3.5 min) offered OPC-61815 and its metabolites relatively high throughput detection, which was greatly beneficial to further clinical bio-sample analysis. The method showed good linearity and sufficient sensitivity in the range of 2.00-1000 ng/mL with a low limit of quantitation (2.00 ng/mL) for each analyte. For samples with concentrations above 1000 ng/mL, 100-fold dilution with blank plasma before sample preparation was accepted. High precision and accuracy, high selectivity and satisfactory recovery of this method were demonstrated. For all of the four analytes, no significant matrix effect or carry-over was observed. The stability of analytes and internal standards under different conditions were evaluated to ensure they were stable during the whole period of storage, preparation and detection. Also, re-injection reproducibility was investigated. In addition, the conversion test showed that almost no OPC-61815 converted into DM-4103 and DM-4107 during sample processing, while attention should be paid to the concentration difference between OPC-61815 and tolvaptan in bioanalysis. The developed UPLC-MS/MS method was successfully applied to an open, single and multiple dose administration phase I trial for monitoring the pharmacokinetics of OPC-61815. This work provided a promising way for further pharmacokinetic study of OPC-61815.

Flexible Resource Allocation-Efficient Water Use Strategies Facilitate Invasion of Invasive Vine Sicyos angulatus L.

The invasive vine Sicyos angulatus L. destroys the natural ecosystem of invaded areas. Understanding the differences in growth and development between S. angulatus and other plants is necessary to explore the invasion mechanisms of S. angulatus and implement appropriate prevention and control measures. Thus, this study compared the growth, photosynthesis, and root characteristics of invasive liana S. angulatus and other three vine plants, Ipomoea nil (L.) Roth, Ipomoea purpurea (L.), and Thladiantha dubia Bunge, at different growth stages: seedling, flowering, and fruiting. The results showed that the total biomass of S. angulatus in the fruiting stage was 3-6 times that of the other three plants, and the root biomass ratio and root-shoot ratio decreased throughout the growth stage. Throughout the growth stage, the total leaf area of S. angulatus was significantly higher than that of the other three plant types, and the specific leaf area of S. angulatus at the seedling and flowering stages was 2.5-3 and 1.4-3 times that of the other three plants, respectively. The photosynthetic rate, stomatal conductance, and transpiration rate of S. angulatus at the fruiting stage were significantly higher than those of the other three plants, and its water use efficiency was higher than that of the other three plants at the three growth stages, indicating its strong photosynthetic capacity. The root activity and root pressure of S. angulatus were also significantly higher than those of the other three plants at the seedling and flowering stages. These results show that S. angulatus flexibly allocates resources to its aboveground parts during the growth stage to ensure that the plant obtains the space necessary for its growth and development and that with the help of higher root pressure and root activity, S. angulatus can maintain higher photosynthesis and water use efficiency with fewer resources. Therefore, the prevention and control of S. angulatus requires a combination of aboveground and underground measures. Spraying conventional weedicide/herbicide and manually removing aboveground plants may lead to its resurgence.

Solobacterium moorei promotes the progression of adenomatous polyps by causing inflammation and disrupting the intestinal barrier.

BACKGROUND: Adenomatous polyps (APs) with inflammation are risk factors for colorectal cancer. However, the role of inflammation-related gut microbiota in promoting the progression of APs is unknown. METHODS: Sequencing of the 16S rRNA gene was conducted to identify characteristic bacteria in AP tissues and normal mucosa. Then, the roles of inflammation-related bacteria were clarified by Spearman correlation analysis. Furthermore, colorectal HT-29 cells, normal colon NCM460 cells, and azoxymethane-treated mice were used to investigate the effects of the characteristic bacteria on progression of APs. RESULTS: The expression levels of inflammation-related markers (diamine oxidase, D-lactate, C-reactive protein, tumor necrosis factor-α, interleukin-6 and interleukin-1ß) were increased, whereas the expression levels of anti-inflammatory factors (interleukin-4 and interleukin-10) were significantly decreased in AP patients as compared to healthy controls. Solobacterium moorei (S. moorei) was enriched in AP tissues and fecal samples, and significantly positively correlated with serum inflammation-related markers. In vitro, S. moorei preferentially attached to HT-29 cells and stimulated cell proliferation and production of pro-inflammatory factors. In vivo, the incidence of intestinal dysplasia was significantly increased in the S. moorei group. Gavage of mice with S. moorei upregulated production of pro-inflammatory factors, suppressed proliferation of CD4+ and CD8+cells, and disrupted the integrity of the intestinal barrier, thereby accelerating progression of APs. CONCLUSIONS: S. moorei accelerated the progression of AP in mice via activation of the NF-κB signaling pathway, chronic low-grade inflammation, and intestinal barrier disruption. Targeted reduction of S. moorei presents a potential strategy to prevent the progression of APs.

NEW PREDICTIVE BIOMARKERS FOR SCREENING COVID-19 PATIENTS WITH RHABDOMYOLYSIS IN COMBINATION WITH CYSTATIN C.

ABSTRACT: Purpose: Cystatin C (CysC) has been linked to the prognosis of corona virus disease 2019 (COVID-19). The study aims to investigate a predictor correlated with CysC screening for poor prognosis in COVID-19 patients combined with skeletal muscle (SKM) impairment and rhabdomyolysis (RM). Methods: A single-center retrospective cohort analysis was carried out. Demographic information, clinical data, laboratory test results, and clinical outcome data were gathered and analyzed. Results: According to the inclusion and exclusion criteria, 382 patients were included in this study. The subjects were divided into three groups based on CysC tertiles. Multivariate analysis revealed that SaO 2 (hazard ratio [HR], 0.946; 95% confidence interval [CI], 0.906-0.987; P = 0.011), CysC (HR, 2.124; 95% CI, 1.223-3.689; P = 0.008), aspartate aminotransferase (AST) (HR, 1.009; 95% CI, 1.000-1.018; P = 0.041), and hypersensitive C-reactive protein (HR, 1.005; 95% CI, 1.000-1.010; P = 0.045) were significantly associated with survivals. The area under curve (AUC) in the model characterized by RM incidence was 0.819 (0.698-0.941), as shown by CysC receiver operating characteristic curves. LDH*CysC and AST*CysC had better predictive values than CysC and the best prediction for RM, with an AUC of 0.880 (0.796,0.964) for LDH*CysC ( P < 0.05, vs CysC) and 0.925 (0.878,0.972) for AST*CysC ( P < 0.05, vs CysC). Conclusion: CysC is an essential evaluation indicator for COVID-19 patients' prognosis. AST*CysC and LDH*CysC have superior predictive value to CysC for SKM, RM, and death, and optimal classification for RM.

Development and validation of a UPLC-MS/MS method for determination of SYHA1807 in a first-in-human study.

Background: SYHA1807 is a novel lysine specific demethylase 1 inhibitor being developed for the treatment of small-cell lung cancer. Aim: This study aimed to establish a ultra-performance liquid chromatography-mass spectrometry (UPLC-MS)/MS method for measuring SYHA1807 in human plasma, supporting its application in a first-in-human study. Methods: SYHA1807 was separated on an ACQUITY UPLC BEH® C18 Column (2.1 × 50 mm, 1.7 µm) after protein precipitation of plasma samples. Mass spectrometry analysis was performed with a Xevo TQS triple quadrupole mass spectrometer utilizing a positive electronic spray ionization source. The established method was fully validated according to bioanalytical guidelines. Results & conclusion: A rapid, specific and robust UPLC-MS/MS method was first established for quantifying SYHA1807 and successfully applied in a first-in-human study.

Safety, tolerability, pharmacokinetics, and pharmacodynamics of single and multiple doses of aficamten in healthy Chinese participants: a randomized, double-blind, placebo-controlled, phase 1 study.

Objectives: Aficamten is a selective, small-molecule allosteric inhibitor of cardiac sarcomere being developed as a chronic oral treatment for patients with symptomatic obstructive hypertrophic cardiomyopathy. This was the first-in-Chinese study aiming to investigate the safety, tolerability, pharmacokinetics, and pharmacodynamics of aficamten in healthy adults. Methods: This double-blind, randomized, placebo-controlled, phase 1 study was conducted in 28 healthy male and female Chinese participants after single ascending dose (SAD) and multi-dose (MD) administrations of aficamten. In the SAD cohort, 16 participants were randomized to receive a single oral dose of aficamten: 10 mg, 20 mg, or placebo. In the MD cohort, 12 participants were randomized to receive multiple doses of aficamten: 5 mg or placebo once daily for 14 days. Safety was monitored throughout the study with electrocardiograms, echocardiograms, clinical laboratory tests, and reporting of adverse events (AEs). Pharmacokinetic profiles of aficamten and metabolites, as well as CYP2D6 genetic impact, were evaluated. Results: A total of 35 treatment-emergent AEs were reported by 14 (50%) participants with mild severity. There were no serious AEs or adverse decreases in left ventricular ejection fraction below 50% during the study. Aficamten was dose-proportional over the dose range of 5-20 mg and accumulated in the MD cohort. Conclusion: Aficamten was safe and well-tolerated in the healthy Chinese adult participants. The pharmacokinetics of aficamten in the Chinese population was comparable to those previously found in Western participants. These phase 1 data support the progression of aficamten into future clinical studies in Chinese patients. Clinical Trial registration: https://clinicaltrials.gov, identifier: NCT04783766.

Predicting the Pharmacokinetics of Orally Administered Drugs across BCS Classes 1-4 by Virtual Bioequivalence Model.

To evaluate the influence of solubility and permeability on the pharmacokinetic prediction performance of orally administered drugs using avirtual bioequivalence (VBE) model, a total of 23 orally administered drugs covering Biopharmaceutics Classification System (BCS) classes 1-4 were selected. A VBE model (i.e., a physiologically based pharmacokinetic model integrated with dissolution data) based on a B2O simulator was applied for pharmacokinetic (PK) prediction in a virtual population. Parameter sensitivity analysis was used for input parameter selection. The predictive performances of PK parameters (i.e., AUC0-t, Cmax, and Tmax), PK profiles, and bioequivalence (BE) results were evaluated using the twofold error, average fold error (AFE), absolute average fold error (AAFE), and BE reassessment metrics. All models successfully simulated the mean PK profiles, with AAFE < 2 and AFE ranging from 0.58 to 1.66. As for the PK parameters, except for the time of peak concentration, Tmax, of isosorbide mononitrate, other simulated PK parameters were all within a twofold error. The simulated PK behaviors were comparable to the observed ones, both for test (T) and reference (R) products, and the simulated T/R arithmetic mean ratios were all within 0.88-1.16 of the observed values. These four evaluation metrics were distributed equally among BCS class 1-4 drugs. The VBE model showed powerful performance to predict the PK behavior of orally administered drugs with various combinations of solubility and permeability, irrespective of the BCS category.

Population pharmacokinetic modeling of ilaprazole in healthy subjects and patients with duodenal ulcer in China.

Aims: This study aimed to develop a population pharmacokinetic (PopPK) model of ilaprazole in healthy subjects and patients with duodenal ulcer in Chinese and investigate the effect of potential covariates on pharmacokinetic (PK) parameters. Methods: Pharmacokinetic data from 4 phase I clinical trials and 1 phase IIa clinical trial of ilaprazole were included in PopPK analysis. Phoenix NLME 8.3 was used to establish a PopPK model and quantify the effects of covariate, such as demographic data, biochemical indicators and disease state on the PK parameters of ilaprazole. The final model was evaluated by goodness-of-fit plots, bootstrap analysis, and visual predictive check. Results: A two-compartment model with first-order elimination successfully described the pharmacokinetic properties of ilaprazole. In the final PopPK model, body weight and sex were identified as statistically significant covariates for volume of peripheral compartment (Vp) and clearance of central compartment (CL), respectively, and disease status was also screened as a significant covariate affecting both CL and Vp. The validation results demonstrated the good predictability of the model, which was accurate and reliable. Conclusion: This is the first population pharmacokinetics study of ilaprazole in the Chinese, and the PopPK model developed in this study is expected to be helpful in providing relevant PK parameters and covariates information for further studies of ilaprazole.

Semi-PBPK Modeling and Simulation to Evaluate the Local and Systemic Pharmacokinetics of OC-01(Varenicline) Nasal Spray.

This study aimed to build a nasal semi-physiologically based pharmacokinetic (PBPK) model to predict the intranasal pharmacokinetic (PK) of the OC-01(varenicline) nasal spray and accelerate the development of this drug. Based on the physiology of the human upper respiratory system, the semi-PBPK model was established and validated using systemic plasma PK data of varenicline previously observed in Americans and Chinese. Drug concentrations, both in respiratory tissue and plasma circulation system, were well simulated, and it was indicated that local concentration at the target site (nasal cavity) was significantly higher than that of plasma when OC-01 nasal spray was administered. The nasal semi-PBPK model successfully depicted the absorption and distribution of intranasal varenicline in the respiratory tissues and provided an alternative to clinical PK study of OC-01 nasal spray in Chinese. Meanwhile the current study presented a viable framework for predicting respiratory concentrations for other novel nasal spray drugs by semi-PBPK modeling.

[Application of MIDD in Clinical Research of Antitumor Drugs].

The antitumor drug has become one of the focused areas in new drug research and development. Their clinical research generally consumes a long period of time, with high cost and high risk. Model-informed drug development (MIDD) integrates and quantitatively analyzes physiological, pharmacological, and disease progression information through modeling and simulation, which can reduce the cost of drug development and improve the efficiency of clinical research. In this essay, Osimertinib and Pembrolizumab are given as examples to illustrate the specific application of MIDD in different phases of clinical research, aiming to provide references for the application of MIDD to guide the clinical research of antitumor drugs.
.

Clinical trials of orphan drugs in China over the decade 2012-2022: Opportunities and challenges.

Rare diseases refer to diseases with very low prevalence. Along with the support of national policies and improvement of research capability, a new landscape for orphan drug is emerging in China. To identity unmet clinical needs and provide insight on the development of orphan drugs, we reviewed the changes over time of orphan drug clinical trials in China from 2012 to 2022. A total of 261 trials of 40 drugs were initiated, of which 66.3% trials were sponsored by Chinese local pharmaceutical enterprises. Among the 261 trials, chemical drugs (about 63.6%) and biological products (35.6%) account for the high proportions, and traditional Chinese medicine (0.8%) was the least; the indications mainly focused on homozygous hypercholesterolemia, hemophilia, multiple sclerosis and idiopathic pulmonary fibrosis; single-arm study design was applied to 50% of the clinical trials, with an average sample size of 52 participants. Additionally, totally 122 trials were completed by January 2022, of which the average duration time was 15.7 months for new drug and 3.5 months for generic drug, respectively. The trends over time illustrated that remarkable progress has been achieved in development of orphan drugs in China since 2012. Given the large patient pool and the rising capability of innovation, it is believed that China will contribute more to the global drug pipelines for rare diseases.

Review of the Chinese Landscape in Phase 1 Clinical Trials for Noncancer Innovative Drugs Over 2015 to 2020.

In recent years, the research and development (R&D) of innovative drugs in China has been dramatically accelerated. And the early clinical study is crucial for drug R&D. However, little is known involving the change of phase I trials for noncancer drugs. We retrieved the data of phase I clinical trials for noncancer innovative drugs on the Registration and Information Disclosure Platform for Drug Clinical Studies on the Center for Drug Evaluation. The number of clinical trials proliferating in recent years and the average annual growth rates of chemical and biological drugs were 55.5% and 42.1%, respectively. Most trials were distributed in Beijing, Shanghai, and other developed coastal cities. Moreover, the clinical trials of innovative drugs in China were focused on the digestive and endocrine systems, whereas the pediatric and orphan drugs were scarce. Based on the data assessment, this work provided comprehensive analysis and suggestions about Chinese drug R&D. Significant advancement has been made in mainland China with the implementation of available policies and the emergence of advanced technologies. Though shortcomings, including uneven geographic distribution and lack of pediatric and orphan drugs, still exist, we believe progress will continue to be made in mainland China.

Development and validation of a high performance liquid chromatography-MS/MS method for determination of SOMCL-15-290 in a first-in-human study.

Background: SOMCL-15-290 is a novel inhibitor that targets FGF receptor, CSF1 receptor and VEGF receptor (kinase insert domain receptor). Aim: This study was aiming at developing a specific high performance liquid chromatography-MS/MS method for quantifying SOMCL-15-290 in human plasma and supporting the first-in-human study. Methods: Plasma samples were prepared using the protein precipitation method and separated on a C18 110A column with acetonitrile and 0.2% formic acid solution as mobile phases. Quantification of SOMCL-15-290 was operated on an Xevo-TQS triple quadrupole tandem mass spectrometer in electrospray ionization positive mode. Results & conclusion: The validated determination method of SOMCL-15-290 has proved feasible and was successfully utilized in the first-in-human study of SOMCL-15-290 in advanced solid tumor patients.

Systemic Review of Animal Models Used in the Study of Crush Syndrome.

ABSTRACT: Crush syndrome (CS), also known as traumatic rhabdomyolysis, is the leading cause of death following extrication from structural collapse due to earthquakes. Due to the unfeasibility of human studies, animal models are used to study crush syndrome pathophysiology, including biochemistry and treatment regimes. The aim of this systematic literature review was to identify the differences and benefits of various animal models used in the study of CS and provide valuable information for design of future research. A systematic search was conducted in two methods: with the filters "(crush syndrome) AND (crush muscle injury)" and with the keywords "(crush syndrome) AND (animal model)" covering all articles in the PubMed databases. The search generated 378 articles. After screening abstracts, 91 articles were retrieved and read, then 11 repeated articles were removed and 2 reference papers were included. We finally reviewed 82 original articles. There appear to be two primary methods employed for inducing crush syndrome in animal models, which are chemically induced injury and physically induced injury. Chemical method mainly includes intramuscular (IM) injection of tissue extract solution and IM injection of 50% glycerine. Physical method can be classified into invasive and non-invasive physical compression by elasticated material, inflatable band and heavy load. Various species of animals have been used to study CS, including mice (13.4%), rats (68.3%), rabbits (11.0%), canines (4.9%), goats (1.2%), and pigs (1.2%). Small animals are suitable for researches exploring the mechanism of disease or drug efficacy while large animals can work better with clinical application-related researches. In regard to the choice of modeling method, compressing the certain muscle of animals by heavy things is superior to others to cause systemic trauma-related rhabdomyolysis signs. In addition, due to the significant burden of crush injuries on animals, further attention shall be paid to the selection of the most suitable anesthetics and appropriate analgesics.

ß1-Blocker improves survival and ventricular remodelling in rats with lethal crush injury.

BACKGROUND: Crush injury/crush syndrome (CI/CS) is the second most common cause of death during earthquakes. Most studies of CI/CS have mainly focused on kidney injury after decompression. Few studies have focused on myocardial injury caused by crush injury and its potential mechanisms. METHODS: We first verified cardiomyocyte injury during compression in rats with a crush injury. The survival rate, electrocardiographic results, histological results, catecholamine changes and cardiac ß1-AR expression were evaluated. Next, we explored the effects of pretreatment with a selective ß1-blocker (bisoprolol) with or without fluid resuscitation on rats with a crush injury. In addition to evaluating the survival rates, biochemical and histological analyses and echocardiographic measurements were also performed. RESULTS: Reduced heart rates, elevated ST segments, and tall-peaked T waves were observed in the rats with a crush injury. The changes in the myocardial enzymes and pathological results demonstrated that myocardial damage occurred during compression in rats with a crush injury. The levels of the catecholamine norepinephrine in both the serum and myocardial tissue were elevated during compression. Pretreatment with a selective ß1-blocker combined with fluid resuscitation significantly improved the survival rates of the rats with lethal crush injury. The myocardial enzymes and pathological results showed that the combined therapy decreased myocardial damage. The echocardiography measurements showed that the rats that received the combined therapy exhibited decreased left ventricular mass (LVM), left ventricular volume at end-systole (LVVs) and left ventricular internal diameter (LVID) compared with the rats with a crush injury. CONCLUSIONS: Our findings demonstrated the presence of myocardial injury in the early stage of compression in rats with a crush injury. Pretreatment with a ß1-blocker (bisoprolol) with fluid resuscitation significantly reduced mortality, decreased myocardial tissue damage, and improved ventricular remodelling in rats with a lethal crush injury.

A Generalized Method for Binary Optimization: Convergence Analysis and Applications.

Binary optimization problems (BOPs) arise naturally in many fields, such as information retrieval, computer vision, and machine learning. Most existing binary optimization methods either use continuous relaxation which can cause large quantization errors, or incorporate a highly specific algorithm that can only be used for particular loss functions. To overcome these difficulties, we propose a novel generalized optimization method, named Alternating Binary Matrix Optimization (ABMO), for solving BOPs. ABMO can handle BOPs with/without orthogonality or linear constraints for a large class of loss functions. ABMO involves rewriting the binary, orthogonality and linear constraints for BOPs as an intersection of two closed sets, then iteratively dividing the original problems into several small optimization problems that can be solved as closed forms. To provide a strict theoretical convergence analysis, we add a sufficiently small perturbation and translate the original problem to an approximated problem whose feasible set is continuous. We not only provide rigorous mathematical proof for the convergence to a stationary and feasible point, but also derive the convergence rate of the proposed algorithm. The promising results obtained from four binary optimization tasks validate the superiority and the generality of ABMO compared with the state-of-the-art methods.

Scaled Simplex Representation for Subspace Clustering.

The self-expressive property of data points, that is, each data point can be linearly represented by the other data points in the same subspace, has proven effective in leading subspace clustering (SC) methods. Most self-expressive methods usually construct a feasible affinity matrix from a coefficient matrix, obtained by solving an optimization problem. However, the negative entries in the coefficient matrix are forced to be positive when constructing the affinity matrix via exponentiation, absolute symmetrization, or squaring operations. This consequently damages the inherent correlations among the data. Besides, the affine constraint used in these methods is not flexible enough for practical applications. To overcome these problems, in this article, we introduce a scaled simplex representation (SSR) for the SC problem. Specifically, the non-negative constraint is used to make the coefficient matrix physically meaningful, and the coefficient vector is constrained to be summed up to a scalar to make it more discriminative. The proposed SSR-based SC (SSRSC) model is reformulated as a linear equality-constrained problem, which is solved efficiently under the alternating direction method of multipliers framework. Experiments on benchmark datasets demonstrate that the proposed SSRSC algorithm is very efficient and outperforms the state-of-the-art SC methods on accuracy. The code can be found at https://github.com/csjunxu/SSRSC.

STAR: A Structure and Texture Aware Retinex Model.

Retinex theory is developed mainly to decompose an image into the illumination and reflectance components by analyzing local image derivatives. In this theory, larger derivatives are attributed to the changes in reflectance, while smaller derivatives are emerged in the smooth illumination. In this paper, we utilize exponentiated local derivatives (with an exponent γ) of an observed image to generate its structure map and texture map. The structure map is produced by been amplified with γ > 1, while the texture map is generated by been shrank with γ < 1. To this end, we design exponential filters for the local derivatives, and present their capability on extracting accurate structure and texture maps, influenced by the choices of exponents γ. The extracted structure and texture maps are employed to regularize the illumination and reflectance components in Retinex decomposition. A novel Structure and Texture Aware Retinex (STAR) model is further proposed for illumination and reflectance decomposition of a single image. We solve the STAR model by an alternating optimization algorithm. Each sub-problem is transformed into a vectorized least squares regression, with closed-form solutions. Comprehensive experiments on commonly tested datasets demonstrate that, the proposed STAR model produce better quantitative and qualitative performance than previous competing methods, on illumination and reflectance decomposition, low-light image enhancement, and color correction. The code is publicly available at https://github.com/csjunxu/STAR.

Multitarget Sparse Latent Regression.

Multitarget regression has recently generated intensive popularity due to its ability to simultaneously solve multiple regression tasks with improved performance, while great challenges stem from jointly exploring inter-target correlations and input-output relationships. In this paper, we propose multitarget sparse latent regression (MSLR) to simultaneously model intrinsic intertarget correlations and complex nonlinear input-output relationships in one single framework. By deploying a structure matrix, the MSLR accomplishes a latent variable model which is able to explicitly encode intertarget correlations via -norm-based sparse learning; the MSLR naturally admits a representer theorem for kernel extension, which enables it to flexibly handle highly complex nonlinear input-output relationships; the MSLR can be solved efficiently by an alternating optimization algorithm with guaranteed convergence, which ensures efficient multitarget regression. Extensive experimental evaluation on both synthetic data and six greatly diverse real-world data sets shows that the proposed MSLR consistently outperforms the state-of-the-art algorithms, which demonstrates its great effectiveness for multivariate prediction.

Multi-Target Regression via Robust Low-Rank Learning.

Multi-target regression has recently regained great popularity due to its capability of simultaneously learning multiple relevant regression tasks and its wide applications in data mining, computer vision and medical image analysis, while great challenges arise from jointly handling inter-target correlations and input-output relationships. In this paper, we propose Multi-layer Multi-target Regression (MMR) which enables simultaneously modeling intrinsic inter-target correlations and nonlinear input-output relationships in a general framework via robust low-rank learning. Specifically, the MMR can explicitly encode inter-target correlations in a structure matrix by matrix elastic nets (MEN); the MMR can work in conjunction with the kernel trick to effectively disentangle highly complex nonlinear input-output relationships; the MMR can be efficiently solved by a new alternating optimization algorithm with guaranteed convergence. The MMR leverages the strength of kernel methods for nonlinear feature learning and the structural advantage of multi-layer learning architectures for inter-target correlation modeling. More importantly, it offers a new multi-layer learning paradigm for multi-target regression which is endowed with high generality, flexibility and expressive ability. Extensive experimental evaluation on 18 diverse real-world datasets demonstrates that our MMR can achieve consistently high performance and outperforms representative state-of-the-art algorithms, which shows its great effectiveness and generality for multivariate prediction.